AI value creation in consumer goods

By Beate Rosenthal, Martin Twesten, Nicolas Wüthrich and Tobias Liebl

Artificial intelligence has moved to the top of the agenda for consumer goods executives. According to our latest AI study, 62% of consumer goods leaders expect major or radical changes to their operating model, but only 38% have started. The window to lead this transformation – rather than react to it – is closing fast. Investment is accelerating, executive sponsorship is at 99% and universal, and 90% of companies have moved well beyond planning. Yet a critical gap persists. ROI does not always live up to expectations: only 31% can demonstrate financial returns, and 14% reach the break-even point on schedule and consistently. The question is no longer whether to invest in AI. It is whether today's approaches are structurally capable of delivering the value that leaders are targeting. This article examines where the consumer goods industry stands, why current models leave substantial value unrealized, and what genuine transformation requires.

Operating model archetypes are defining today's landscape and differences in ROI

Our research identified four distinct AI operating model archetypes among consumer goods companies, with different maturity and ROI achievement level: target-driven, enabler-driven, governance-driven, and capability-driven. These archetypes differ meaningfully in AI maturity and ambition. Enabler-driven models, built on strong data foundations and focused on applying AI to solve concrete business challenges, consistently outperform the others in terms of measurable ROI achievement. Understanding which archetype your organization represents — and what it would take to shift — is the starting point for any credible AI value creation strategy.

Best practices emerge but are defining a ceiling to be overcome

We found five principles that summarize how best in class consumer goods companies are leaning forward on the AI journey. Companies that consistently follow these principles generate more value than those that do not.

1. Orchestrate, don’t oursource: the most AI fluent companies build up capabilities in a way that they orchestrate the tech architecture between inhouse and external development, for instance by inhousing decisive capabilities in functions e.g. AI in innovation which is a key business lever in cosmetics. They select what’s needed to deliver better and faster on business strategy. They build and enhance required future capabilities.
2. Rewire, don’t wrap: means getting the fundamentals done. That applies both to data and infrastructure as well as to actively shaping skills and culture. Having done those fundamentals in previous digitalization actions can generate advantage.
3. Platform the risk: AI leaders take governance seriously and design it in without strangling the organizations. This can include ethical standards as well as creating business modeling templates for governing AI use cases.
4. Federate, don’t silo: Experience shows that the best ideas don’t always come from the center. Accordingly, those companies that encourage experimentation and guide scaling are more successful than those that are focussing too early on central use cases. Portfolio companies we’ve seen intentionally spread experimentation across business units.
5. Operate don’t just launch: once go life is completed many walk away, in our best practice cases, teams consider go life the base line for continuous improvement. Once scaled, they go back to the original pilot and improve it and the teams capabilities to apply AI at the highest standard.

While many master fundamentals, we have a strong belief that it’s not enough to optimize archetypes to take a leap in AI value creation – some companies will choose to reimagine the future towards AI-First.

Here is the structural ceiling. Best practices that optimize existing operating models are, by definition, constrained by those models and processes. Role-based structures, legacy processes, and governance frameworks designed for human-coordinated work cannot fully absorb the capabilities of autonomous, outcome-driven AI systems. Applying AI within legacy structures produces incremental gains. It does not produce the step-change performance that current and near-term AI technology makes possible. However, the single largest barrier to AI scaling across all archetypes is not technology – it is people: adopting and building AI literacy, shifting mindset, and embedding new ways of working is needed.

From incremental improvement to AI-First reimagination

There are several drivers why companies need to move to the next level.

Technology is moving fast with Agentic AI becoming the new baseline, not an experiment. Leading organizations are moving beyond task automation toward autonomous, context-aware systems that execute and coordinate work end-to-end. Higher levels of value and change require an operating model shift. Real value comes from rethinking how decisions are made, how teams collaborate, and how workflows across the enterprise. Legacy processes and structures limit AI’s potential. Unlocking value means questioning existing roles, handovers and governance and replacing them with outcome-driven, AI-orchestrated workflows. Companies that continue to apply AI within existing structures will reach a performance ceiling. Those that redesign their operating models around AI-orchestrated, outcome-led workflows access a fundamentally different level of capability.

An AI-First organization is not defined by the volume of AI tools deployed. It is defined by whether AI is embedded in the core business model, whether processes are designed around outcomes rather than roles, and whether governance is built for autonomous execution rather than human handovers.

Productivity implications of this shift can be material

Our analysis estimates that an AI-First operating model can unlock 50%+ total productivity gains at the enterprise level. Specific domains show particularly strong potential: customer service and after sales (40–60%), finance and procurement (40–60%), applications and platforms (40–50%), and sales, marketing, and pricing (30–50%). Supply chain and logistics, manufacturing, and corporate functions including HR and legal also show meaningful gains in the range of 20–40%. These figures represent the upper bound of what becomes accessible when organizations move from optimizing individual tasks to redesigning end-to-end processes around AI-orchestrated outcomes.

AI-First operating models in FMCG are rare and emerging, mainly across start-up landscape or where speed and agility is at the core of the business model. In one example, the company has a long-term focus on outperforming and out-innovating industry processes and speed end to end. That translates into AI at the core supply chain with real-time demand sensing from store-level sales data driving production decisions within 2–3-week cycles and AI-based inventory allocation.

Two distinct paths forward

We suggest two transformation paths available to consumer goods leaders, each suited to different organizational contexts and levels of readiness. The first path, Acceleration, focuses on uplifting AI value within the current operating model. This means syncing AI and business strategy, applying a rigorous portfolio logic to use case deployment, building scalable data and technology foundations, and federating AI governance through an AI Officer model with distributed ownership. The second path, Re.imagination, is more fundamental. It begins with a zero-based rethinking of the target operating model—setting aside legacy structures and backward-engineering a roadmap from a future-state vision built around AI-orchestrated outcomes. This can either start in a specific function or business unit and be scaled accordingly.

Both paths are viable, and the right choice depends on a company's current archetype, strategic ambition, and organizational readiness. What our research makes clear is that standing still is not a stable position. The competitive advantage in consumer goods will increasingly be defined by how rapidly companies master adoption of AI technology into their core operations or challenge existing operating and business models — not by how many AI projects they have launched.

Register now for our newsletter and get regular insights into Consumer Goods & Retail topics.